What Cosmos composite indexes should I make to optimize a Gremlin query with "And" steps?

Question

I'm not sure if Gremlin queries even leverage indexes when traversing. If they do, I'm not sure what indexes I should make for a query like so:

 g.V().hasLabel("Item")
  .and(
    __.has("partitionKey","F6369365-31A9-4035-93C1-9873FECFC4E8"),
    __.has("facilityId","B67EAF94-92D9-4137-035E-08DB668F9820")
  )
  .and(
    __.has("effectiveStartDate",lte(202307051320169088)),
    __.has("effectiveEndDate",gt(202307051320169088))
  )

Should each "and" step be made into a composite index or should I have one long composite index? Perhaps, I should make one that includes /label, /partitionKey, and /facilityId then a second that has /effectiveStartDate and /effectiveEndDate.

I can't find any documentation on how best to create indexes in Cosmos for Gremlin queries. Any insight will be helpful. Thanks!